Results

cesm2.ssp245

diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error ks_mean_on_coarse_res_with_extremes qqplot_mae acf_mae extremogram_mae
lstm.cesm2.ssp245 1.03% 0.946 0.768 0.190 0.186 0.056 0.087 0.033
xgboost.cesm2.ssp245 1.53% 0.909 0.609 0.154 0.347 0.093 0.140 0.074
nv.cesm2.ssp245 1.57% 0.914 0.641 0.173 0.323 0.103 0.126 0.065
cnn.cesm2.ssp245 -1.62% 0.949 0.740 0.138 0.259 0.087 0.100 0.054

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram

cesm2.ssp370

diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error ks_mean_on_coarse_res_with_extremes qqplot_mae acf_mae extremogram_mae
lstm.cesm2.ssp370 0.53% 0.952 0.772 0.198 0.142 0.057 0.078 0.013
nv.cesm2.ssp370 0.72% 0.909 0.641 0.174 0.245 0.115 0.118 0.044
xgboost.cesm2.ssp370 0.76% 0.899 0.609 0.152 0.324 0.106 0.130 0.041
cnn.cesm2.ssp370 -2.06% 0.957 0.748 0.145 0.194 0.105 0.092 0.023

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram

cesm2.ssp585

diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error ks_mean_on_coarse_res_with_extremes qqplot_mae acf_mae extremogram_mae
lstm.cesm2.ssp585 0.97% 0.952 0.783 0.201 0.130 0.052 0.068 0.038
cnn.cesm2.ssp585 -1.56% 0.960 0.763 0.161 0.147 0.084 0.082 0.045
xgboost.cesm2.ssp585 1.65% 0.891 0.610 0.146 0.310 0.107 0.122 0.071
nv.cesm2.ssp585 1.69% 0.897 0.643 0.163 0.279 0.118 0.111 0.066

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram

ec_earth3.ssp434

diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error ks_mean_on_coarse_res_with_extremes qqplot_mae acf_mae extremogram_mae
lstm.ec_earth3.ssp434 0.92% 0.954 0.766 0.195 0.257 0.101 0.085 0.094
cnn.ec_earth3.ssp434 -1.54% 0.968 0.743 0.137 0.273 0.080 0.098 0.098
xgboost.ec_earth3.ssp434 -4.33% 0.971 0.638 0.180 0.405 0.173 0.132 0.135
nv.ec_earth3.ssp434 -4.51% 0.990 0.665 0.135 0.388 0.178 0.128 0.131

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram

mri_esm2_0.ssp245

diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error ks_mean_on_coarse_res_with_extremes qqplot_mae acf_mae extremogram_mae
lstm.mri_esm2_0.ssp245 0.81% 0.959 0.764 0.202 0.194 0.065 0.083 0.030
cnn.mri_esm2_0.ssp245 -1.51% 0.970 0.750 0.163 0.224 0.066 0.097 0.036
xgboost.mri_esm2_0.ssp245 36.02% 0.546 0.368 0.269 0.373 1.425 0.128 0.045
nv.mri_esm2_0.ssp245 36.48% 0.610 0.522 0.239 0.233 1.442 0.084 0.026

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram

mri_esm2_0.ssp370

diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error ks_mean_on_coarse_res_with_extremes qqplot_mae acf_mae extremogram_mae
lstm.mri_esm2_0.ssp370 0.91% 0.949 0.769 0.183 0.144 0.057 0.065 0.011
cnn.mri_esm2_0.ssp370 -1.38% 0.960 0.756 0.156 0.183 0.073 0.079 0.017
xgboost.mri_esm2_0.ssp370 36.67% 0.529 0.361 0.273 0.323 1.451 0.115 0.032
nv.mri_esm2_0.ssp370 37.29% 0.599 0.526 0.213 0.204 1.474 0.067 0.019

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram

mri_esm2_0.ssp434

diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error ks_mean_on_coarse_res_with_extremes qqplot_mae acf_mae extremogram_mae
lstm.mri_esm2_0.ssp434 0.43% 0.948 0.767 0.205 0.170 0.061 0.078 0.014
cnn.mri_esm2_0.ssp434 -1.96% 0.954 0.748 0.151 0.189 0.090 0.094 0.014
xgboost.mri_esm2_0.ssp434 35.90% 0.543 0.371 0.278 0.357 1.420 0.125 0.027
nv.mri_esm2_0.ssp434 36.37% 0.607 0.525 0.227 0.221 1.438 0.084 0.027

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram